1. Field of the Invention
The present invention relates to an optical signal processor and, more particularly, to an improved photoelectric transducer.
2. Description of the Related Art
FIG. 1(a) is a plan view of a photosensor cell which serves as a photoelectric transducer, while FIG. 1(b) is a sectional view taken along the line A--A' in FIG. 1(a), and FIGS. 2a and 2b and equivalent circuits of the photosensor cell. In these figures, the same members or portions or those which are equivalent to each other are denoted by the same reference numerals.
Although the photosensor cell shown in FIGS. 1a and 1b is arranged so that picture elements which are adjacent to each other are lined up in columns and rows, it is, of course, possible to dispose these picture elements so that the elements in each row are offset from their counterparts in adjacent rows (i.e., interpolation arrangement) in order to increase resolution in the horizontal direction.
The structure of this photosensor cell is as follows:
As shown in FIGS. 1(a) and 1(b), various layers, regions, etc., such as those mentioned below are formed on an n-type silicon subtrate 1 to form the photosensor cell: a passivation layer 2; an insulating oxide layer 3 constituted by a silicon oxide layer; element-isolating regions 4 constituted by an insulating layer or a polysilicon layer for electrically insulating photosensor cells which are adjacent to each other; an n.sup.- -region 5 of low impurity concentration which is formed by the epitaxial technique or other similar means; a p-region 6 which constitutes the base of a bipolar transistor; an n.sup.+ -region 7 which constitutes the emitter of the bipolar transistor; an electrode 8 formed of an electrically conductive material, e.g., aluminum (Al), for reading out a signal to an external device; a capacitor electrode 9 facing the p-region 6 across the insulating layer 3 and employed to apply pulses to the p-region 6 which is in a floating state; an electrode 10 connected to the capacitor electrode 9; an n.sup.+ -region 11 formed on the reverse surface of the substrate 1 for forming an ohmic contact; and an electrode 12 for applying a potential to the collector of the bipolar transistor.
Referring to FIG. 2a, which shows an equivalent circuit of the above-described photosensor cell, a capacitor Cox13 has an MOS (metal oxide semiconductor) structure which is constituted, as shown in FIG. 1b, by the electrode 9, the insulating layer 3 and the p-region 6, and a bipolar transistor 14 is constituted, as shown in FIG. 1b, by the n.sup.+ -region 7 which serves as the emitter, the p-region 6 as the base, and the n.sup.- -region 5 and the region 1 as the collector. As will be clear from these figures, the p-region 6 is a floating region.
The bipolar transistor 14 may be represented by an equivalent circuit which is shown in FIG. 2b and is constituted by a base-emitter junction capacitance Cbe15, a base-emitter pn junction diode Dbe16, a base-collector junction capacitance Cbc17, a base-collector pn junction diode Dbc18 and current sources 19 and 20.
The following is a description of the basic operation of the photosensor cell having the above-described structure.
The basic operation of the photosensor cell includes a charge-accumulating operation during irradiation with light, a reading operation and a refreshing operation. In the charge-accumulating operation, for example, the emitter is grounded through the electrode 8, and the collector is biased to a positive potential through the electrode 12. The base is assumed to be reverse-biased with respect to the emitter.
In this state, when light 20 infringes upon the photosensor cell from the obverse side thereof, as shown in FIG. 1(b), electron-hole pairs are generated within the semiconductor. In the electron-hole pairs, the electrons flow out toward the n-region 1 since it is biased to a positive potential, but the holes are gradually accumulated in the p-region 6. The accumulation of holes in the p-region 6 causes the potential of the p-region 6 to change gradually toward positive. At this time, the potential Vp generated at the p-region 6 (which is the base) by the accumulation of the holes excited by the incident light may be given by Vp=Q/C, where Q represents the amount of charges (holes in this case) accumulated in the base, and C represents the total junction capacitance of the capacitances Cbe15 and Cbc17.
It is to be noted here that, as the cell size is reduced as a result of an increase in resolution of the device, the quantity of light incident on each photosensor cell is reduced, and the amount of accumulated charge Q is also decreased; however, since the junction capacitance also decreases in proportion to the reduction in cell size, the potential Vp generated by the incidence of light is maintained at a substantially constant level. This is because the photosensor cell has an extremely simple structure, as shown in FIG. 1, and this makes it possible to obtain a remarkably large effective light-sensitive or light-receiving area.
The following is a description of the operation of reading out the voltage generated by the charges thus accumulated in the p-region 6.
In a reading operation, the emitter and the electrode 8 are left floating, and the collector is held at a positive potential Vcc.
Assuming now that the potential at the p-region or base 6 when biased to a negative potential before the irradiation with light is represented by -Vb, and the accumulated charge voltage generated by the irradiation is represented by Vp, the base potential at the time when a reading operation is to be effected is -Vb+Vp. When, in this state, a positive voltage Vr for reading is applied to the electrode 9 through the electrode 10, this positive voltage Vr is capacitance-divided by the oxide layer capacitance Cox13, the base-emitter junction capacitance Cbe15 and the base-collector junction capacitance Cbc17, and the base potential becomes as follows: ##EQU1## If this base potential is forward-biased excessively by Vbs which is represented by the following equation: ##EQU2## then, the base potential is forward-biased more than the accumulated charge voltage Vp generated by the irradiation with light. In consequence, electrons are injected into the base from the emitter, and since the collector potential is positive, the injected electrons are accelerated by the drift field so as to reach the collector.
FIG. 3(a) is a graph showing the relationship between the read voltage and the accumulated charge voltage Vp when Vbs=0.6 V.
As will be understood from the graph, if the read time (a period of time during which the read voltage Vr is being applied to the capacitor electrode 9) is set at about 100 nsec (nanoseconds) or more, the linear relationship between the accumulated charge voltage Vp and the read voltage is ensured within a range of about four decimal places, and high-speed reading operation is possible. In the above-described calculation example, the capacitance of the electrode 8 is 4 pF (picofarads), and the total junction capacitance Cbe+Cbc is 0.014 pF, which means the capacitance ratio is about 300:1. In this case, however, the voltage Vp generated in the p-region or base 6 is not attenuated at all, and a reading operation effected at extremely high speed is made possible by the effect of the bias voltage Vbs. This is because the amplifying function of the photosensor cell works effectively. Thus, since the output voltage is favorably large, the level of fixed pattern noise and that of random noise attributable to the output capacity become small relatively, and it is therefore possible to obtain a signal of remarkably excellent S/N (signal to noise) ratio.
FIG. 3(b) is a graph showing the relationship between the read time, the bias voltage Vbs, and the linear relationship between the accumulated charge voltage and the read voltage which has been described above: namely, excellent linearity on the order of four decimal places can be obtained within a high-speed read time of about 100 nsec when the bias voltage Vbs is set at 0.6 V.
It is possible, according to the graph shown in FIG. 3(b), to ascertain the read time required for the read voltage which depends on the bias voltage Vbs to reach a desired ratio (%) with respect to the accumulated charge voltage. Accordingly, when the read time and the required linearity are determined from the overall design of a particular image sensor, the required bias voltage Vbs can be determined using the graph shown in FIG. 3(b).
Another advantage of the photosensor cell having the above-described arrangement is that it is possible to read out the holes accumulated in the p-region or base 6 in a non-destructive manner because of the extremely small probability of the electrons and holes in the p-region 6 recombining with each other. This means that it is possible to provide a novel function in terms of the system operation when the photosensor cell of the above-described arrangement is employed to constitute an image sensor.
The accumulated charge voltage Vp can be held in the p-region or base 6 for a remarkably long period of time, but the maximum holding time is limited by the dark current which is thermally generated in a depletion layer at the junction. In the above-described photosensor cell, however, the depletion layer extends in the region 5 which has an extemely low impurity concentration; therefore, the number of electron-hole pairs thermally generated therein is favorably small by virtue of its excellent crystallizability.
The operation of refreshing the p-region 6 by destroying the charges already accumulated therein will be described below.
In a photosensor cell of the above-described arrangement, the charges accumulated in the p-region 6 do not disappear even when a reading operation has been completed, as already described above. In order to input new optical information, therefore, it is necessary to carry out a refreshing operation for the purpose of causing the charges which have been accumulated to disappear. At the same time, the p-region 6, which is left floating, must be charged so that the potential thereof is a predetermined negative voltage.
With the photosensor cell arranged as described above, the refreshing operation is effected by applying a positive voltage to the electrode 9 through the electrode 10 in the same manner as in the case of the reading operation. At this time, the emitter is grounded through the electrode 8, while the collector is held at either ground potential or a positive potential through the electrode 12. FIG. 4(a) shows an equivalent circuit for analyzing the refreshing operation, the circuit comprising an example in which the collector side is grounded.
When, in this state, a positive voltage Vrh is applied to the electrode 9, the voltage shown below is momentarily applied to the base 6 in a manner similar to that in the aforementioned reading operation as a result of capacitance division effected by the oxide layer capacitance Cox13, the base-emitter junction capacitance Cbe15 and the base-collector junction capacitance Cbc17: ##EQU3## This voltage causes the base-emitter junction diode Dbe16 and the base-collector junction diode Dbc18 to be forward-biased and thereby made conductive, so that a current starts to flow, and the base potential gradually lowers.
FIG. 4(b) shows the results of calculation carried out with respect to changes in the potential of the base 22 which is in a floating state at the time of refreshing operation, as one example of the dependence of the base potential on time. The axis of abscissa of the graph shown in FIG. 4(b) represents the refreshing time, that is, the elapse of time from the moment the refreshing voltage Vrh is applied to the electrode 9, while the axis of ordinate represents the base potential, and the initial base potential is employed as a parameter. The initial base potential is a potential at the floating base which is measured when the refreshing voltage Vrh is applied to the electrode 9 and which is determined by Vrh, Cox, Cbe, Cbc and the charges which have been accumulated in the base.
It will be understood from FIG. 4(b) that the base potential invariably lowers along one straight line on the semi-logarithmic graph after a certain period of time has elapsed irrespective of the level of the initial base potential.
There are two methods or operations of removing any positive voltage at the p-region 6 by applying a positive voltage through the MOS capacitor Cox for a certain period of time and thereby charging the p-region 6 to a negative potential. One of them is an operation in which negative charges are accumulated in the p-region 6 as the result of the holes having positive charges flowing out from the p-region 6 into the n-region 1 which is mainly held in a grounded state.
The other operation of accumulating negative charges in the p-region 6 is effected by the electrons which flow out from the n.sup.+ -region 7 and the region 1 into the p-region 6 and recombine with the holes therein.
A solid-state image sensor which is constituted by photosensor cells having the arrangement described above involves two modes, that is, a complete refreshing mode in which the base potentials of all the cells are brought to zero voltage by a refreshing operation (in this case, 10.sup.0 sec, i.e., one second is needed in the example shown in FIG. 4(b)), and a transient refreshing mode in which the variable component of the base potential which depends upon the accumulated charge voltage Vp disappears although a predetermined voltage remains (in this case, a refreshing pulse of 10 .mu.sec (microseconds) to 10.sup.0 sec, i.e., one second is generated in the example shown in FIG. 4(b)).
Whether the image sensor is to be operated in the complete refreshing mode or in the transient refreshing mode is determined in accordance with the purpose of use of the device.
The basic operation of the photosensor cell in accordance with the above-described arrangement is thus constituted by the charge accumulating operation effected by irradiation with light, the reading operation and the refreshing operation. Accordingly, it is possible to observe incident light or read out optical information with each of the operations employed as a basic cycle.
As has been described above, the basic structure of the above-described photosensor cell is extremely simple compared with those of the image sensors which are described in Japanese Patent Laid-Open specifications Nos. 150878/1981, 157073/1981 and 165473/1981, respectively. It is therefore possible for the photosensor cell to satisfactorily cope with any need to increase resolution. In addition, the photosensor cell preserves the merits of the devices disclosed in those specifications which are offered by virtue of the amplifying function that is an excellent feature thereof, such as a lowered level of noise, an increased output, a widened dynamic range, a non-destructive reading operation, etc.